Seeded planetary bearing fault in a helicopter gearbox—A case study

Abstract

Helicopters have been extensively employed as a versatile tool in modern society, thus guaranteeing operation safety and flightworthiness has always been our top-tier concerns. However, many accidents indicate that current helicopter health and usage monitoring system (HUMS) are not accurate and effective enough to fulfil this goal, especially for detecting an imminent planetary bearing fault (CAA Authority 2006). To study this case and address such issues, an experiment was undertaken with a special rig, comprising of real helicopter main gearbox (MGB) and dynamometer as a variable load. Defects of different sizes were seeded at the outer race of bearings inside second epicyclic planetary module. Vibration signals of different test conditions (input speed, load and defect sizes) were recorded for post-processing. Measured vibration data contains overwhelming background noise and planetary gear meshes, making it cumbersome to identify the bearing defect. Various signal processing techniques are applied to identify the bearing defect, including signal pre-whitening, gear/bearing signal separation, kurtogram, envelope analysis, high-order spectral analysis and so on. The comparative results suggest that certain combinations of aforementioned methods are more effective than single process techniques. One case of successful detection is demonstrated in this paper. This study potentially benefits improving HUMS effectiveness and supports helicopter maintenance decision-making.